Mastering 'Metrics: The Path from Cause to Effect

Product Information

The five most valuable econometric methods, or what the authors call the Furious Five--random assignment, regression, instrumental variables, regression discontinuity designs, and differences in differences--are illustrated through well-crafted real-world examples (vetted for awesomeness by Kung Fu Panda's Jade Palace). Does health insurance make you healthier? Randomized experiments provide answers. Are expensive private colleges and selective public high schools better than more pedestrian institutions? Regression analysis and a regression discontinuity design reveal the surprising truth. When private banks teeter, and depositors take their money and run, should central banks step in to save them? Differences-in-differences analysis of a Depression-era banking crisis offers a response. Could arresting O. J. Simpson have saved his ex-wife's life? Instrumental variables methods instruct law enforcement authorities in how best to respond to domestic abuse. Data scientists, on the other hand, don't often think about economics at all. From their perspective the two disciplines have basically no overlap. So they struggle to see why they should care about what an economist has to say about anything. This is primarily driven by the popular misperception of economics being about business questions. Imagine their frustration when economists start telling them that their results are wrong.

I would be hard pressed to name another econometrics book that can be read for enjoyment yet provides useful quantitative insights."— Financial Analysts Journal This valuable book connects the dots between mathematical formulas, statistical methods, and real-world policy analysis. Reading it is like overhearing a conversation between two grumpy old men who happen to be economists—and I mean this in the best way possible."—Andrew Gelman, Columbia University Angrist, JD, and J-S Pischke (2015), Mastering Metrics: The Path from Cause to Effect, Princeton University Press.In terms of the chapters itself, I think they are very topical and will cover a lot of the modern research; the book pulls away from a fundamental issue - no matter what the methods are, the thought of comparison and counterfactuals is not emphasized enough I feel. Consider a standard econometrics textbook - say Wooldridge - it actually draws a framework where you know - no matter what the empirical problem is, you need to think in terms of identification, endogeneity and the underlying logic of counter-factuals. They certainly bring in a lot of that - where they talk about apples-to-apples comparison; but the emphasis is not approached as a general method of empirical analysis and the book can go far if that is emphasized. Thus in terms of binding the various methods - (a) a comparison and (b) a generalized empirical strategy might help get the econometrics logic through to a wider audience. Modern econometrics is more than just a set of statistical tools—causal inference in the social sciences requires a careful, inquisitive mindset. Mastering 'Metrics is an engaging, fun, and highly accessible guide to the paradigm of causal inference."—David Deming, Harvard University Economists view data scientists as regression monkeys (probably the worst insult you can give someone in economics). When they look at data science they just see extremely elaborate efforts at curve fitting. Since economists don't think curve fitting is all that interesting or useful for doing economics, they scoff at neural networks and boosting. Imagine their horror when they see data science moving into their territory.

The first chapter Randomized Trials outlines basic experimental concepts like treatment, outcome, control and treatment group, the fundamental problem that we can always only observe one reality in one person, and the idea that randomization makes "other things equal" (p. xii). It also points out why perfect randomization is difficult to achieve in real life. Furthermore, the issue of statistical significance in the interpretation of results is discussed, as analyses are usually only based on samples drawn from populations. Around five years ago, Joshua D. Angrist and Jörn-Steffen Pischke published their first joint book on econometrics tools for causal inference: Mostly harmless econometrics (2009). Although this book is excellent in many regards (e.g., more than 5000 quotes on Google Scholar), it was not as harmless as the title might suggest. Mastering 'Metrics: The path from cause to effect now fills this gap, as it is a truly nontechnical introduction. As already introduced in the first chapter, treatment and control groups are not necessarily equal in all other aspects, especially under non-randomized conditions. Therefore, the idea of "Regression" is discussed in the next chapter. Regression is presented as a conditioning technique that only delivers credible results if all variables that introduce group differences apart from the treatment are observed. Such variables are then computationally made equal across the groups, so that causal inference can be made. The authors emphasize that, in most natural settings, selection bias can have multiple sources that are usually not all observable. In such cases, the power of regression is limited. The five most valuable econometric methods, or what the authors call the Furious Five—random assignment, regression, instrumental variables, regression discontinuity designs, and differences in differences—are illustrated through well-crafted real-world examples (vetted for awesomeness by Kung Fu Panda’s Jade Palace). Does health insurance make you healthier? Randomized experiments provide answers. Are expensive private colleges and selective public high schools better than more pedestrian institutions? Regression analysis and a regression discontinuity design reveal the surprising truth. When private banks teeter, and depositors take their money and run, should central banks step in to save them? Differences-in-differences analysis of a Depression-era banking crisis offers a response. Could arresting O. J. Simpson have saved his ex-wife’s life? Instrumental variables methods instruct law enforcement authorities in how best to respond to domestic abuse. First, the content. Mastering 'Metrics does a pretty good job of covering the intuition (and some of the math) behind random assignment, regression, instrumental variables, regression discontinuity designs, and difference in differences. I think their treatment of these topics would be most useful to someone who was trying to read modern applied econometrics (or political science). Ideally the reader would have taken enough statistics that they can focus on trying to grasp the concept of potential outcomes rather than trying to work through the algebra. The methods that are covered are extremely important in social science and so having an idea of what they do and why we use them is helpful.

Pragmatism

Written by true 'masters of 'metrics,' this book is perfect for those who wish to study this important subject. Using real-world examples and only elementary statistics, Angrist and Pischke convey the central methods of causal inference with clarity and wit."—Hal Varian, chief economist at Google

Another relevant factor with the book is that the passages do not lead you to read on - rather they are too academic! If the intention is for a wider audience and for a more diversified crowd, then the importance of leading readers onto the next issues is of supreme importance. For eg: they are discussing an issue and then the next issue comes up as a next section. There is no sense of direction as to why am I reading about an issue and where do the connections matter - in terms of comprehending the entire topic, the reader is left on his own. Personally I found the extended metaphor that econometrics is kung fu to be annoying. I think the authors believed that they were making the material more accessible by treating it less reverently, which I agree could have been an effective communication strategy, but I think it mostly fell flat. If I'm cringing at your puns I'm not learning about local average treatment effects. Moreover, I think the metaphor that econometrics is kung fu is actually harmful. Kung fu is mysterious and mystical. It's studied at the feet of a master over the course of a lifetime. The master might have you wash floors for a year, without offering a reason. There is definitely an art to econometrics, but clouding econometrics in mysticism does more to protect the reputation of the teacher than it does to advance the student's learning. Others may disagree but this grasshopper would have preferred we spend less time in the dojo and more time in the computer lab. The fact that there are not endless instrumental variables given in all areas of interest, often makes it necessary to use other approaches like Differencesin-Differ enees, which is illustrated in chapter 5. The authors explain how developments of control and treat- ment groups can indicate treatment effects, even in the absence of randomization. The approach assumes that even if groups differ in the outcome from the very beginning, a non-parallel development of the groups can be attributed to the treatment, which is again illustrated clearly using econometric examples. Hamermesh, DS (2013), “Six Decades of Top Economics Publishing: Who and How?”, Journal of Economic Literature, 162-172.

References

With humor and rigor, this book explores key approaches in applied econometrics. The authors present accessible, interesting examples—using data-heavy figures and graphic-style comics—to teach practitioners the intuition and statistical understanding they need to become masters of 'metrics. A must-read for anyone using data to investigate questions of causality!"—Melissa S. Kearney, University of Maryland and the Brookings Institution Our focus on five core econometric tools is a natural consequence of contemporary econometric practice, which owes little to the formalities of the classical linear regression model, the arcane statistical assumptions of generalised least squares, or the elaborate simultaneous equations framework that fill so many texts. We begin with randomised trials, which set our standard for research validity, moving on to a detailed but model-free discussion of regression, the tool most likely to be used by practitioners. Our regression application — estimating the effects of private college attendance on later earnings — shows the power of regression to turn night into day when it comes to causal conclusions. Wielding econometric tools with skill and confidence, Mastering ‘Metrics uses data and statistics to illuminate the path from cause to effect. You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer Few fields of statistical inquiry have seen faster progress over the last several decades than causal inference. With an engaging, insightful style, Angrist and Pischke catch readers up on five powerful methods in this area. If you seek to make causal inferences, or understand those made by others, you will want to read this book as soon as possible."—Gary King, Harvard University

The disconnect between econometric teaching and econometric practice goes beyond questions of tone and illustration. The most disturbing gap here is conceptual. The ascendance of the five core econometric tools – experiments, matching and regression methods, instrumental variables, differences-in-differences and regression discontinuity designs – marks a paradigm shift in empirical economics. In the past, empirical research focused on the estimation of models, presented as tests of economic theories or simply because modelling is what econometrics was thought to be about. Contemporary applied research asks focussed questions about economic forces and economic policy.If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about. So i have almost reached halfway chapter 4 where RDD is being discussed. I found the chapters imbalanced. Like the IV chapter was very heavy and was not a smoother flow like the other ones. Angrist, Joshua D. & Pischke, Jörn-Steffen (2015). Mastering 'Metrics: The path from cause to effect. Princeton, Oxford: Princeton University Press, 304 p., 35 USD, ISBN 978-0-691-15284-4 The Regression Discontinuity Designs are depicted in chapter 4 and distinguished from the instrumental variables approach. The fact that variables in here have a fixed cutoff point - resulting from an external rule - which either completely determines how a treatment manifests or increases its likelihood, is illustrated. Individuals close to this cut-off can be seen as equal in other characteristics. For example, Angrist and Pischke investigate whether young adults die more often on their 21st birthday. The regression discontinuity in the mortality rate around the birthday is then interpreted as an indicator for the effect of the minimum legal drinking age, defined by law ("Some young people appear to pay the ultimate price for the privilege of downing a legal drink", p. 164). The basic idea why this method is also a robust path to causal inference is explicitly discussed.